San Francisco Bay Area
My research interests intersect Formal Methods, Cryptography, and Human-Computer Interaction. I aim to design and develop intuitive verification tools that are easy to integrate into existing pipelines, with a focus on cryptographic applications. My current work centers on advancing SMT solvers for zero-knowledge proof verification.
Working with Professor Michael Coblenz on developing tools and domain specific programming languages for oceanographers
Project: Regex + Advised by: Nadia Polikarpova and Taylor Berg-Kirkpatrick. Working on nuero-symbolic regular expression synthesis from only positive examples. Preliminary results presented at SYNT workshop and submitted to a student competition at POPL. Project: Functional Debugging Advised by: Michael Coblenz Investigated how Haskell and functional programmers in general debug. Results submitted to PL+HCI workshop. Project: Print and Probability Advised by: Taylor Berg-Kirkpatrick Contributed to developing state of the art neural network model for early modern printer attribution. Results accepted to AAAI
Researched, extended, united and automated verification checks done by Data Science, Business Intelli- gence, and Data Engineering teams for the market metadata optimization campaign. Build a database uniting metadata from 5 different providers, 6 products and 3 marketplaces.
Developed an intelligent robotic automation process for customer discovery for the PRT team. Investigated key text features in Form 5500 pdf files indicating the future sale of contract. Created CPU-parallelized pipeline for web scraping, OCR, and keyword identification. Led seminars on web scraping and OCR.
Investigated and compared ways to optimize the performance of a Deep Transfer Learning Model for image classification in different ML Frameworks: Keras, TensorFlow and PyTorch. Results were used to teach ML researchers at San Diego Super Computer’s Cyberinfrastructure-Enabled Machine Learning (CIML) Summer Institute.